BFG EX series

[Minerva]

High ripple on both EX1000 and EX1200 models. Efficiency is good but too bad they mucked up the output. Back to the drawing board, Andyson!

If you value your system you should not use this supply! Ripple is a silent, deadly killer and these supplies EXCEED ATX spec on 12V!!! :|

I sent my 1200 back to Newegg and bought the Galaxy EVO 1250 from zipzoomfly instead ($229 after MIR) .

If they can fix this ripple this would be a good supply! It's compact and quiet. If you really NEED 1kW+ of power however I'd get something else though because it runs hotter than hell. 80 deg C air out the back means the internal temps must be ready to burn something. I'd really be scared of it blowing up personally with that much load.

 

[jonnyGURU]

I think you're probably talking about the Oklahoma Wolf review over at jonnyGURU.com.

First off, Wolf only reviewed the EX-1000, not the EX-1200.

Furthermore, only Wolf found the high ripple on the +3.3V and +5V and we still don't know why he's seeing it (we've been discussing it in the mod forum at jg.com).

Neither Andyson nor myself can duplicate the ripple he had seen and even Paul from HardOCP has a unit that he's in the process of reviewing and he can't reproduce the high ripple either.

The +12V does not exceed spec until test 5 and only on the modular PCIe connectors. Keep in mind that test 5 put an 18A load on a single PCIe connector. A 6-pin PCIe connector is only designed to deliver up to 6.25A. An 8-pin is designed to deliver 12.5A. So only if you had a build where you could fully load a 6-pin and 8-pin connector, you would hit 19A and possibly see 126mV of ripple (spec being 120mV).

I was able to reproduce is the high ripple on the modular PCIe connectors and that's already been addressed on production units by using larger capacitors on the back of the modular PCB.

So I'm sorry to hear that you went through the trouble of RMA-ing your PSU based on one review when there's very likely absolutely nothing wrong with it and I'm sorry it got you so wound up over his review that you felt the need to post this kind of thread on Anandtech. I'm guilty of knee jerk reactions too, but you should also allow yourself to step back and really assess the situation.

 

[Minerva]

Normally I would not sound alarm on OW's review of the 1kW unit however when I saw the OC3D review here not just the high ripple but the high temps really concern me! I'm not trying to diss BFG or Andyson but it seems that there is a problem in Houston guys.

I like the size of the PSU and it's nearly silent but if I'm getting lethal doses of ripple to my SSD's I don't want to think about that. :Q

 

[HOOfan 1]

Originally posted by: jonnyGURU
....

thought you had abandoned anandtech.

 

[jonnyGURU]

Originally posted by: Minerva
Normally I would not sound alarm on OW's review of the 1kW unit however when I saw the OC3D review here not just the high ripple but the high temps really concern me! I'm not trying to diss BFG or Andyson but it seems that there is a problem in Houston guys.

One needs to think logically about PSU temps as well. The temperature a PSU produces is in direct correlation to the amount of power the unit is putting out and the efficiency of the unit. The unit is efficient, so there shouldn't be any unusual amount of heat coming from the unit. The actual Andyson branded unit (MT9, it's called) actually has a fan that doesn't even spin up until internal temps hit 50°C. I didn't feel comfortable with this and opted for a fan that ramped up gradually. I'm not sure where the reviewers took their temperatures from, but clearly the readings were done at a particular "hot spot" on the unit. But if the unit actually ran significantly hotter than any other unit on the market, common sense would dictate that the unit would also be less efficient... and it's not.

Bottom line: I'm satisfied with the EX and this is why I had BFG bring it to market. I was aware of Wolf's finding prior to his review, but do not believe in sensoring him for them. Instead, I preferred to use his findings as "input" and attempt to duplicate them because ANY product on the market has room for improvement and I value data from multiple sources even if they vary drastically from my own.

Originally posted by: Minerva
I like the size of the PSU and it's nearly silent but if I'm getting lethal doses of ripple to my SSD's I don't want to think about that. :Q

Your SSD's only use +5V, so ripple on a +12V PCIe modular connector is sort of a moot point.

Originally posted by: HOOfan 1
thought you had abandoned anandtech.

Just been too damn busy... then I saw this post and.. well....

 

[Minerva]

Originally posted by: jonnyGURU


One needs to think logically about PSU temps as well. The temperature a PSU produces is in direct correlation to the amount of power the unit is putting out and the efficiency of the unit. The unit is efficient, so there shouldn't be any unusual amount of heat coming from the unit. The actual Andyson branded unit (MT9, it's called) actually has a fan that doesn't even spin up until internal temps hit 50°C. I didn't feel comfortable with this and opted for a fan that ramped up gradually. I'm not sure where the reviewers took their temperatures from, but clearly the readings were done at a particular "hot spot" on the unit. But if the unit actually ran significantly hotter than any other unit on the market, common sense would dictate that the unit would also be less efficient... and it's not.

Bottom line: I'm satisfied with the EX and this is why I had BFG bring it to market. I was aware of Wolf's finding prior to his review, but do not believe in sensoring him for them. Instead, I preferred to use his findings as "input" and attempt to duplicate them because ANY product on the market has room for improvement and I value data from multiple sources even if they vary drastically from my own.

Where did I or the reviewer debate efficiency? You want to know what the problem is? Form factor. Every other PSU in this class (1200W) is much longer. You're dealing with the immutable laws of physics here. Simply put the heat sinks are too small and run extremely hot. You're basically saying the reviewers picked the hottest point on the back. That's laughable because they measured ambient air IN and air discharged subtracting the smaller value from the larger which the difference is the delta. This supply has an extremely high delta which indicates the dwell time for the cooling medium (air) is too long for the surface area to protect internal components. (the heatsink is too small!). So what you really have here is a duty cycle rated product. Electrically it's capable of producing 1.2kW of power but thermally it cannot sustain that output continuously without breaking down. Fortunately in this industry this can happen because 98% of the pc users would never have a 1.2kW DC load on a PC. Those that did would not be happy. This is further evidenced by the 40 degrees Centigrade rating vs. everyone else at 50 degrees Centigrade.

At full power given 84% efficiency the components must dissipate nearly 200W of heat! I'd be willing to bet if that box was loaded 100% the internal heatsink would sizzle if you spat on it! :laugh: The best CPU HSF are huge for a reason! Imagine if you took an i7 975 at 1.5V and 4.5GHz and mounted that heatsink on it!

Originally posted by: jonnyGURU

Your SSD's only use +5V, so ripple on a +12V PCIe modular connector is sort of a moot point.

I'm using a Super Micro M28E2 SAS mobile rack which takes a pair of connectors using 12VDC. I really don't want this (along with $14k in drives!) being fed dirty power. The original chassis was using an Ultra X3 1600W that performs flawlessly but was too long to fit in this smaller box. That idea has been scrapped. There's truly no substitute for cubic inches!

The decision to return it was not impulsive by any means. A lot of thought went into this. The ripple was the icing on the cake.

 

[jonnyGURU]

I didn't say you disputed the efficiency. I merely asked you to consider it. What happens to the AC power that's not converted to DC? It's converted to heat. What's not converted to heat is converted to what? DC power. So if the PSU is putting out the same amount of DC power as another PSU and at the same or better efficiency, where is this supposed extra heat coming from?

 

[Minerva]

The reviews are not measuring heat output (typically in btu/hr) but the delta from air in to air out. If you increase airflow with the box producing the same amount of heat the delta will be lower. If you use a 130CFM Delta screamer running WOT at the highest levels the air would be much cooler. (and you could probably dry laundry too! ) My X3 1600 definitely gets very noticeable at the upper levels (100A+ 12V!) and the air is very warm but definitely not as hot. Easiest way to bring the temps down is set the ramping steeper on the controller providing the fan can deliver the goods.

My brother said he would've gladly taken the supply off my hands and cracked the seal for a look (he's an EE in training and of course he would being he isn't paying for it!)

 

[jonnyGURU]

Originally posted by: Minerva
The reviews are not measuring heat output (typically in btu/hr) but the delta from air in to air out. If you increase airflow with the box producing the same amount of heat the delta will be lower. If you use a 130CFM Delta screamer running WOT at the highest levels the air would be much cooler.

Exactly. You also don't have as much air escaping back into the case.

Like I said, the MT9 was designed for a fan that doesn't even spin up until 50°C. I opted for a gradual ramp up. The PSU is engineered to handle these temps.

BTW: High nipple?

 

[SunSamurai]

"Watch out for high nipple!"


YES MAM!